100
L. G.-Weil et al.
5.2.1 Sampling Properties of Remote Sensors
There are important differences between the sampling scheme of multi-spectral radiometers and radar altimeters. The first are imaging systems providing a synoptic
view of the ocean surface. Those used in this work (e.g. the Advanced Very High Resolution Radiometer, AVHRR; the Moderate Resolution Imaging Spectroradiometer,
MODIS; and the Sea-viewing Wide Field-of-view Sensor, SeaWiFS) have a spatial
resolution on the order of ∼ 1 km at nadir, and a viewing swath width larger than
1500 km. The spatial resolution degrades toward the edge of the swath, such that in
many applications, as in this study, these parts of the scene are not used. The second,
being non-imaging radars, only provide a profile of sea surface height below the
satellite along their ground tracks. Then, interpolation is needed to build SLA maps.
In the present case, SLA maps were generated by the French Archiving, Validation and Interpretation of Satellite Oceanographic (AVISO) data center, combining
the measurements of several radar altimeters in order to minimize errors due to the
paucity of data of one single altimeter missions in the composition of SLA maps
(Le Traon and Dibarboure 1999). Specifically, in this work it was used the reference delayed time product (SSALTO/DUACS User Handbook 2012). This merges
in one map the measurements of two altimeter missions, on the same two orbits (first
Topex/Poseidon and ERS followed by Jason-1 and Envisat). Thus it ensures a homogeneous and stable sampling along the available time period. It has been shown that
Gaussian-shaped eddies with e-folding scales of at least 0.4
◦ are resolved in these
maps (Chelton et al. 2011).
With respect to temporal sampling, during the years analyzed in more detail in
this work (i.e. the 1998–1999 period), there were usually 4 AVHRR-recorded SST
(or BT) scenes available per day. This high-frequency rate is due to the fact that
infrared sensors can observe the Earth during day and night, and that in 1998–1999
there were two satellites, with an AVHRR on board, observing the area of study
at different times. Conversely, just 1 Chl-a scene per day was available, because
visible sensors operate during daytime only, and in 1998–1999 SeaWiFS was the only
one available to observe ocean colour from space. The actual frequency of useful
SST and Chl-a images is lower than the nominal one mentioned above, because
no information can be retrieved by infrared and visible radiometers when clouds
cover the ocean, a limitation that renders the time interval between consecutive
useful SST or Chl-a scenes highly irregular. On the other hand, radar altimeters
measure a microwave signal going through clouds, but this signal is degraded if
rain is present. The Canary Islands area is placed in a dry subtropical zone and
therefore altimeter data from this region are not seriously affected by rain (Tran et al.
2005). Although atmospheric effects are less a problem in the microwave region,
the altimeter sampling characteristics generate a very poor temporal resolution (in
excess of 10 days). By merging data from more than one altimeter in the same data
product, AVISO produces regularly one SLA map of any given area every 7 days.
L. G.-Weil et al.
5.2.1 Sampling Properties of Remote Sensors
There are important differences between the sampling scheme of multi-spectral radiometers and radar altimeters. The first are imaging systems providing a synoptic
view of the ocean surface. Those used in this work (e.g. the Advanced Very High Resolution Radiometer, AVHRR; the Moderate Resolution Imaging Spectroradiometer,
MODIS; and the Sea-viewing Wide Field-of-view Sensor, SeaWiFS) have a spatial
resolution on the order of ∼ 1 km at nadir, and a viewing swath width larger than
1500 km. The spatial resolution degrades toward the edge of the swath, such that in
many applications, as in this study, these parts of the scene are not used. The second,
being non-imaging radars, only provide a profile of sea surface height below the
satellite along their ground tracks. Then, interpolation is needed to build SLA maps.
In the present case, SLA maps were generated by the French Archiving, Validation and Interpretation of Satellite Oceanographic (AVISO) data center, combining
the measurements of several radar altimeters in order to minimize errors due to the
paucity of data of one single altimeter missions in the composition of SLA maps
(Le Traon and Dibarboure 1999). Specifically, in this work it was used the reference delayed time product (SSALTO/DUACS User Handbook 2012). This merges
in one map the measurements of two altimeter missions, on the same two orbits (first
Topex/Poseidon and ERS followed by Jason-1 and Envisat). Thus it ensures a homogeneous and stable sampling along the available time period. It has been shown that
Gaussian-shaped eddies with e-folding scales of at least 0.4
◦ are resolved in these
maps (Chelton et al. 2011).
With respect to temporal sampling, during the years analyzed in more detail in
this work (i.e. the 1998–1999 period), there were usually 4 AVHRR-recorded SST
(or BT) scenes available per day. This high-frequency rate is due to the fact that
infrared sensors can observe the Earth during day and night, and that in 1998–1999
there were two satellites, with an AVHRR on board, observing the area of study
at different times. Conversely, just 1 Chl-a scene per day was available, because
visible sensors operate during daytime only, and in 1998–1999 SeaWiFS was the only
one available to observe ocean colour from space. The actual frequency of useful
SST and Chl-a images is lower than the nominal one mentioned above, because
no information can be retrieved by infrared and visible radiometers when clouds
cover the ocean, a limitation that renders the time interval between consecutive
useful SST or Chl-a scenes highly irregular. On the other hand, radar altimeters
measure a microwave signal going through clouds, but this signal is degraded if
rain is present. The Canary Islands area is placed in a dry subtropical zone and
therefore altimeter data from this region are not seriously affected by rain (Tran et al.
2005). Although atmospheric effects are less a problem in the microwave region,
the altimeter sampling characteristics generate a very poor temporal resolution (in
excess of 10 days). By merging data from more than one altimeter in the same data
product, AVISO produces regularly one SLA map of any given area every 7 days.